Antitumor activity of a homing peptide that targetstumor lymphatics and tumor cellsPirjo Laakkonen*†, Maria E. Åkerman*‡, Hector Biliran*§, Meng Yang¶, Fernando Ferrer*, Terhi Karpanen†,Robert M. Hoffman¶, and Erkki Ruoslahti*�

*The Burnham Institute, La Jolla, CA 92037; †Molecular�Cancer Biology Laboratory, Biomedicum Helsinki, University of Helsinki, P.O.B. 63 (Haartmaninkatu8), FIN-00014 Helsinki, Finland; ‡Department of Bioengineering, University of California at San Diego, La Jolla, CA 92093; and ¶AntiCancer, Inc.,San Diego, CA 92111

Contributed by Erkki Ruoslahti, May 10, 2004

LyP-1 is a peptide selected from a phage-displayed peptide librarythat specifically binds to tumor and endothelial cells of tumorlymphatics in certain tumors. Fluorescein-conjugated LyP-1 and arelated peptide, LyP-1b, strongly accumulated in primary MDA-MB-435 breast cancer xenografts and their metastases from i.v.peptide injections, allowing visualization of orthotopic tumors inintact mice. The LyP peptide accumulation coincided with hypoxicareas in tumors. LyP-1 induced cell death in cultured human breastcarcinoma cells that bind and internalize the peptide. Melanomacells that do not bind LyP-1 were unaffected. Systemic LyP-1peptide treatment of mice with xenografted tumors induced withthe breast cancer cells inhibited tumor growth. The treated tumorscontained foci of apoptotic cells and were essentially devoid oflymphatics. These results reveal an unexpected antitumor effect bythe LyP-1 peptide that seems to be dependent on a proapoptotic�cytotoxic activity of the peptide. As LyP-1 affects the poorlyvascularized tumor compartment, it may complement treatmentsdirected at tumor blood vessels.

phage display � tumor targeting � live imaging � therapy

Tumor blood vessels express molecular markers that distin-guish them from normal blood vessels. Many of these tumor

vessel markers are related to angiogenesis, but some are selectivefor certain tumors (1). Markers that distinguish the vasculatureof tumors at the premalignant stage from the vasculature of fullymalignant tumors in the same tumor system have also beendescribed (2, 3). Recent data from our laboratory indicate thatlymphatic vessels in tumors are also specialized, because a cyclic9-amino acid peptide, LyP-1, binds to the lymphatic vessels incertain tumors, but not to the lymphatics of normal tissues (4).

The lymphatic system is an important route of tumor metas-tasis. Many cancers preferentially spread through the lymphatics.Recent discoveries of growth factors and molecular markers forlymphatic endothelial cells have made possible detailed studiesof the relationship of tumor cells and the lymphatic vasculatureof tumors (5–9). The use of marker proteins such as LYVE-1 (6),podoplanin (5), and Prox-1 (10) has shown that lymphatic vesselsare abundant in the periphery of tumors and that many tumorsalso contain lymphatics within the tumor mass (4, 11). However,the intratumoral lymphatic vessels are generally not functional intransporting tissue fluid (12) and are often filled with tumor cells(4, 13). Recent experimental and clinical data strongly suggestthat the number of lymphatics in a tumor, perhaps their size aswell, and the expression of lymphangiogenic growth factors areimportant determinants in the ability of a tumor to metastasize(14–18).

Thus, it may become possible to reduce metastasis by specif-ically targeting tumor lymphatics (and the tumor tissue adjacentto these vessels) for destruction. The LyP-1 peptide, whichspecifically binds to tumor lymphatics (4), provides one potentialavenue for developing reagents that can specifically destroytumor lymphatics. This peptide also binds to the tumor cells in

tumors that contain LyP-1-positive lymphatics, further expand-ing the potential of this peptide.

We show here that i.v. injected LyP-1 strongly and specificallyaccumulates in breast cancer xenografts over time, localizingpreferentially in hypoxic areas. We also report that LyP-1 has aproapoptotic�cytotoxic effect on tumor cells and that systemicadministration of the LyP-1 peptide inhibits breast cancer xeno-graft growth in mice. The treated tumors contain foci ofapoptotic cells and reduced numbers of lymphatic vessels. Thesefindings suggest that LyP-1 may provide a starting point for thedevelopment of new antitumor agents.

Materials and MethodsCell Lines and Tumors. MDA-MB-435 human breast carcinomacells and C8161 human melanoma cells were maintained inDMEM supplemented with 10% FCS. Nude BALB�c nu�numice were injected s.c. or into the mammary fat pad with 1 � 106

tumor cells to induce tumors. A vascular endothelial growthfactor (VEGF)-C-transfected MDA-MB-435 cell line was pre-pared as previously reported for MCF7 cells (13).

Antibodies and Immunohistology. Blood vessels were visualized bystaining tissue sections with monoclonal antibodies againstCD-31, CD-34, or MECA-32 (all rat anti-mouse antibodies fromPharmingen). A polyclonal rabbit anti-mouse LYVE-1 antibody(4) and a rat monoclonal anti-mouse podoplanin antibody(provided by Kari Alitalo, University of Helsinki) were used tovisualize lymphatic vessels. The primary antibodies were de-tected with goat anti-rabbit or anti-rat Alexa 594 (MolecularProbes).

Biodistribution of fluorescein-conjugated peptides was exam-ined after i.v. injection (100 �l of 1 mM peptide solution in 200�l of PBS) into the tail vein of a mouse. The peptide was allowedto circulate for various periods of time, and the mouse wasperfused through the heart with 4% paraformaldehyde. Tissueswere removed, soaked in 30% sucrose in PBS overnight, andfrozen in OCT embedding medium (Tissue-Tek). Alternatively,tumor-bearing mice were i.v. injected with 500 �l of 1 mMfluorescein-conjugated peptide in PBS, and the peptide wasallowed to circulate for 16–20 h.

The whole-body imaging was done under a blue light, by using theimaging system of a fluorescence stereo microscope (model LZ12;Leica, Deerfield, IL) equipped with a mercury 50-W lamp (19).

Determination of Vessel Density in Tissues. Frozen tumor sectionswere stained with antibodies against CD-34 and podoplanin (5)to visualize the tumor-associated blood and lymphatic vessels.Using �200 magnification, each microscopic field in the hori-

Abbreviation: VEGF, vascular endothelial growth factor.

§Present address: Department of Pathology, School of Medicine, Wayne State University,540 East Canfield Road, Detroit, MI 48201.

�To whom correspondence should be addressed. E-mail: ruoslahti@burnham.org.

© 2004 by The National Academy of Sciences of the USA

www.pnas.org�cgi�doi�10.1073�pnas.0403317101 PNAS � June 22, 2004 � vol. 101 � no. 25 � 9381–9386

MED

ICA

LSC

IEN

CES

Dow

nloa

ded

by g

uest

on

July

11,

202

0

zontal and the vertical directions was counted for the presenceof the two types of vessels.

Hypoxia. Hypoxic areas in the tumor were visualized by i.v.injection of a hypoxia marker 2-nitroimidazole (EF5) (20) intotumor-bearing mice (10 �l of 10 mM EF5 per g), followed byCy3-conjugated mouse anti-EF5 (provided by Randall S.Johnson, University of California at San Diego). CulturedMDA-MB-435 cells were grown on coverslips and incubatedovernight at 37°C to allow for attachment and spreading of thecells. Half of the cells were transferred to a hypoxia chamber(0.1% oxygen�5% CO2) and incubated overnight under hy-poxic conditions. Fluorescein-conjugated peptides (10 �M)were added to the cells in 1% BSA in DMEM and incubatedfor 3 h, followed by fixation with 4% paraformaldehyde in PBS.The coverslips were mounted on glass slides by using Vecta-Shield mounting media with 4�,6-diamidino-2-phenylindole(Vector Laboratories).

Cytotoxicity Assay. Cytotoxic efficacy of the different peptideswas judged by measuring the release of a cytoplasmic enzyme,lactate dehydrogenase, from damaged cells into the superna-tant by using a colorimetric assay Cytotoxicity Detection Kit(LDH assay; Roche Diagnostics). MDA-MB-435 cells wereplated on 96-well plates (6,000 cells per well) and incubatedovernight at 37°C to allow for attachment and spreading of thecells. Cells were washed once with PBS, and 50 �l of 2% BSAin DMEM was added to the cells. Peptides were added in50 �l of H2O and incubated for 24–72 h at 37°C. After theincubation, the cells were spun down (1,000 rpm, 10 min), andthe supernatant was transferred to a new plate. The colorreaction was added to the cells and incubated for 25 min beforethe absorbance was read at 492 nm. Cells incubated with 50 �lof H2O and 50 �l of 2% BSA in DMEM served as a backgroundcontrol, and cells incubated with 1% Nonidet P-40 showed themaximal cytotoxic value. The cytotoxicity was determined asa percentage of the maximal value after the subtraction of thebackground.

Tumor Treatment Studies. Tumor-bearing mice were treated withi.v. injections of peptides beginning 4 weeks after tumor cellinoculation. The injections were administered twice a week for4–5 weeks. Tumor volumes were measured once a week andwere calculated according to the formula V � width �height � depth�2, derived from the formula for the volume ofan ellipsoid (21). Student’s t test was used for statisticalanalysis of the results. The animal experiments reported herewere approved by The Burnham Institute Animal ResearchCommittee.

Synthesis of Fluorescein-Conjugated Peptides. Peptides were syn-thesized by using Fmoc-protected amino acids (Nova Bio-chem) and HATU (PE Biosystems, Foster City, CA) as acoupling reagent in dimethylformamide activated with diisio-propylethylamine. All peptides were amide-capped at the Cterminus by the use of Fmoc-PAL-PEG-PS resin (PE Biosys-tems). The peptides were conjugated with f luorescein at the Nterminus by reacting with f luorescein isothiocyanate isomer(FITC, Aldrich) in dimethylformamide for 20 h in the presenceof diisiopropylethylamine.

ResultsFluorescein-Conjugated Homing Peptides Accumulate in Tumor Tissue.Intravenously injected LyP-1 peptide was observed to home totumor-associated lymphatic vessels and tumor cells in MDA-MB-435 xenografts and some other tumors (4). In this earlierwork, the peptide was allowed to circulate for �20 min. Tooptimize the accumulation of LyP-1 in these tumors, we

studied the distribution of the peptide for longer periods oftime. We found striking accumulation of f luorescein-conjugated LyP-1 in tumors several hours after the injection.At 16–20 h, the tumors of the LyP-1-injected mice werebrightly f luorescent in whole-body f luorescent imaging (19) ofintact mice (Fig. 1A). Tumors from the mice injected with acontrol peptide showed no f luorescence (Fig. 1B). Imaging ofdissected tumors and organs from the same animals revealedstrong f luorescence in tumors from mice injected with LyP-1,whereas no f luorescence was detectable in the tumors from thecontrol peptide-injected mice. Other tissues showed no spe-cific f luorescence with either peptide (Fig. 1 C and D).

We then confirmed the tumor-specificity of LyP-1 by quan-tifying the fluorescence in the dissected tissues. Tumor fluores-cence from the control peptide injection was too low to beaccurately distinguished from the background, but LyP-1 con-centration was at least 15- to 40-fold higher in the tumors thanthat of the control peptide, whereas fluorescence in other tissueswas not significantly different from the background (Fig. 1E). Apeptide closely related to LyP-1 (CNKRTRGGC; H.B., J. A.Hoffman, P.L., and E.R., unpublished data) also strongly accu-mulated in the tumors (LyP-1b in Fig. 1E). These results showthat the LyP-1 peptides accumulate in the MDA-MB-435 tumorswith extraordinary efficiency and that the accumulation isspecific. The LyP-1 fluorescence was mainly present in tumor cellnuclei (Fig. 1F), whereas the control peptide was essentiallynegative in tumor tissue (Fig. 1G). No fluorescence was detectedin other tissues with any of the peptides (shown for LyP-1 in braintissue in Fig. 1H).

LyP-1 Recognizes Metastatic Lesions. MDA-MB-435 tumor cellstransfected with the lymphangiogenic growth factor VEGF-Cproduce tumors with increased number of lymphatic vesselsand enhanced propensity to metastasize into regional lymphnodes and the lungs (13, 16, 22). In agreement with the abilityof LyP-1 to recognize tumor lymphatics, LyP-1 accumulationin the VEGF-C-expressing tumors seemed to be stronger thanin the parental-line tumors (data not shown). LyP-1 peptidealso homed to the metastatic lymph nodes of the MDA-MB-435�VEGF-C tumor mice (Fig. 2 A–E), colocalizing withlymphatic endothelial markers (arrows in Fig. 2C) and tumorcells (Fig. 2E) within the lymph nodes. No LyP-1 f luorescencewas detected in the vessels of normal lymph nodes; the nucleiof a few isolated cells that appeared to be leukocytes werepositive (Fig. 2B Inset). Metastatic foci in lungs were alsopositive for Lyp-1 (Fig. 2F). These results show that metastasescan retain the LyP-1 binding of the primary tumor and that thesame tumor can induce the LyP-1-binding epitope in thelymphatic vessels of more than one tissue.

LyP-1 Peptide Recognizes Hypoxic Areas in Tumors. The tumor cellsthat accumulated LyP-1 formed clusters within the tumors, andthese clusters contained few blood vessels (Fig. 3A) but werepositive for lymphatic endothelial markers (Fig. 3B). The LyP-1-positive tumor cell clusters were strikingly similar to clusters oftumor cells revealed by uptake of hypoxia markers (23). Thissimilarity, and the lack of blood vessels, led us to examine apossible connection between LyP-1 binding and hypoxia. Intra-venously injected hypoxia reagent EF5 and fluorescein-labeledLyP-1 accumulated in the same areas in the tumors, but thestaining for the two markers seemed to be mutually exclusive atthe level of individual cells (Fig. 3 C and D). If EF5 was injectedfirst, the homing of LyP-1 was reduced (Fig. 3C) and vice versa(Fig. 3D). Moreover, injecting LyP-1 or EF5 alone gave astronger tumor signal for both compounds than the coinjections.In contrast, the accumulation of EF5 in C8161 melanomaxenografts, which are not recognized by LyP-1 (4), was unaf-fected by coinjecting LyP-1 (data not shown). These results

9382 � www.pnas.org�cgi�doi�10.1073�pnas.0403317101 Laakkonen et al.

Dow

nloa

ded

by g

uest

on

July

11,

202

0

indicate that LyP-1 preferentially localizes in hypoxic parts oftumors and that LyP-1 and EF5 specifically affect one another’srecognition of hypoxic tumor cells.

Serum Starvation Increases Binding of Fluorescein-Conjugated LyP-1to Cultured MDA-MB-435 Cells. We next sought to reproduce theeffect of hypoxia on tumor cell recognition in vitro. We cultured

MDA-MB-435 cells under hypoxic conditions but detected noincrease in the number of cells that were positive for fluorescein-conjugated LyP-1 (data not shown). However, we did see anincrease in the number of cells that had taken up LyP-1 when wemaintained the cells in low serum (compare Fig. 3 E and F).Counting of LyP-1-positive cells showed that the difference was2.5-fold. These results suggest that LyP-1 homing to tumors may

Fig. 1. Specific accumulation of lymphatic homing peptides in tumors. Mice bearing orthotopic MDA-MB-435 xenograft tumors were i.v. injected withfluorescein-conjugated LyP-1 (A) or a fluorescein-conjugated control peptide (ARALPSQRSR) (B). The mice were anesthetized 16–20 h later and examined forfluorescence under blue light. Tumor fluorescence of a LyP-1-injected tumor mouse is shown in A. No fluorescence was detected in tumors of mice injected withthe control peptide (B). After the external examination, the mice were killed, and tumor, kidneys, spleen, and liver were excised and examined for fluorescence.LyP-1 produced intense fluorescence in the tumor, whereas no fluorescence was detectable in other organs (C). Even when imaged directly, no fluorescence wasobserved in the control peptide-injected tumor (D). The gallbladder is autofluorescent and appears as a green spot in C and D. E shows quantification of theimaging results for LyP-1 and for LyP-1b, which was analyzed in similar experiments. Mice that did not receive any fluorescent compound were used to determinethe level of autofluorescence in tissues, and this background was subtracted from the experimental values. The graph shows a representative experiment of three.(F–H) Mice injected with peptides as in A and B were perfused through the heart, and their tumors were examined microscopically. Strong LyP-1 fluorescenceis seen in the nuclei (visualized by 4�,6-diamidino-2-phenylindole staining) of tumor cells (F). No appreciable fluorescence from the control peptide is seen in tumortissue (G), and all normal tissues tested were negative for all peptides (the result for LyP-1 in the brain is shown in H). T, tumor; B, brain; H, heart; Lu, lungs; Li,liver; S, spleen; K, kidneys. Magnification: F and G, �100; H, �200.

Laakkonen et al. PNAS � June 22, 2004 � vol. 101 � no. 25 � 9383

MED

ICA

LSC

IEN

CES

Dow

nloa

ded

by g

uest

on

July

11,

202

0

not be directly related to hypoxia but may result from theattendant nutrient starvation.

LyP-1 Binding and Internalization Induce Cell Death. Studying theinternalization of the fluorescein-conjugated LyP-1 peptide incultured cells, we noticed that the LyP-1 positive cells tended toround up, and the morphology of their nuclei frequently sug-gested apoptosis. To investigate whether LyP-1 caused cell death,we incubated MDA-MB-435 cells with unlabeled LyP-1 andmonitored cell lysis. Incubation with LyP-1 resulted in a con-centration-dependent increase in cell lysis with an IC50 of �66�M (Fig. 4A). C8161 human melanoma cells, which do not bindLyP-1 (4), were not affected by the peptide. Control peptidesthat resemble LyP-1 in their amino acid composition and�orcyclic structure (CRVRTRSGC, Fig. 4A) and two other peptides[CGEKRTRGC, a variant of LyP-1, which has no cell-bindingactivity (4); and KECQSRLLSCP, (data not shown)] had noeffect on the viability of either cell line. Thus, LyP-1 specificallykills cells that bind this peptide.

Systemic Treatment with LyP-1 Inhibits Tumor Growth and Reducesthe Number of Tumor Lymphatics. Given that LyP-1 had an in vitrocytotoxic effect on the MDA-MB-435 tumor cells, we examined

the effect of LyP-1 on tumor growth in vivo. We gave MDA-MB-435 or MDA-MB-435�VEGF-C tumor mice biweekly i.v.injections of the LyP-1 peptide, starting after the mice hadestablished palpable tumors. Fig. 4B shows one of three similartreatment experiments. The LyP-1 peptide inhibited tumorgrowth formed by both cell lines. The average reduction of tumorvolume relative to the control-treated mice was �50% and highlysignificant (P � 0.005). Increasing the dose of the LyP-1 peptidedid not improve the efficacy of the compound (data not shown).The tumors of the LyP-1-treated animals contained numerousTUNEL-positive cells, indicating apoptosis, whereas little apo-ptosis was detected in the tumors of the control-treated mice(Fig. 4 C and D). The increased apoptosis in the LyP-1 group wasspecific for the tumor tissue; other tissues did not containsignificant numbers of TUNEL-positive cells (data not shown).

LyP-1-treatment selectively reduced the number of lymphaticvessels in the tumors, while having a less prominent effect on theblood vessel density in the same tumors (Fig. 5). These results arein agreement with the in vivo homing pattern of the fluorescein-conjugated LyP-1 peptide to the lymphatics in MDA-MB-435tumor-bearing mice (4). It seems that the lymphatic endothelialcells in the tumor are also susceptible to LyP-1.

Fig. 2. Lymphatic homing peptide recognizes metastases of VEGF-C-expressing tumors. Fluorescein-conjugated LyP-1 peptide was i.v. injected intomice bearing orthotopic VEGF-C-expressing MDA-MB-435 tumors and al-lowed to circulate for 15 min. The tumor, lymph nodes, lungs, kidneys, andliver were removed and prepared for immunohistology. Lymphatic vessels inlymph node metastases (A–D) were visualized by staining with anti-LYVE-1antibodies followed by goat anti-rabbit Alexa 594 (red, A and C). Nuclei werevisualized by 4�,6-diamidino-2-phenylindole staining (blue, B and D). A and Band C and D show the same microscopic fields with different staining. LyP-1peptide (green) is present in the nuclei of cells in and around enlargedlymphatic vessels in lymph node metastases. These cells are tumor cells asjudged by their intense staining with anti-VEGF-C antibody (red, E). Thepeptide is also seen in the nuclei of lymphatic endothelial cells (arrows in C).No LyP-1 accumulated in a tumor-free lymph node (B Inset). A metastatic lungtumor also accumulates LyP-1 (F; LyP-1, green; nuclei, blue). Magnification,�200; Inset, �50.

Fig. 3. LyP-1 peptide recognizes cell clusters that lack blood vessels butcontain lymphatics. Fluorescein-conjugated LyP-1 peptide was i.v. injectedinto MDA-MB-435 tumor-bearing mice and allowed to circulate for 15 min.LyP-1 peptide was seen in cell clusters throughout the tumor (green, A and B).These areas did not contain blood vessels, as judged by staining with the bloodvessel endothelial marker, MECA-32 (red, A) but were often positive for thelymphatic endothelial markers, LYVE-1 (red, B) and podoplanin (not shown).The hypoxia marker EF5 (red), injected 8 h before LyP-1 (green), localized inthe LyP-1-positive patches within the tumors (C). Reversing the order of theinjections reduced the amount of EF5 in the LyP-1-positive patches (D). Thepresence of the two compounds at the cellular level seemed to be mutuallyexclusive. (E and F) LyP-1 binding to cultured cells is increased by serumstarvation. Fluorescein-conjugated LyP-1 peptide was added to MDA-MB-435cells cultured either in 10% (E) or 0.1% (F) serum, and the binding and uptakeof the peptide by the cells was determined 3 h later. Serum starvationincreased the number of LyP-1-positive cells (LyP-1, green; nuclei, blue).Magnification: A–D, �200; E and F, �100.

9384 � www.pnas.org�cgi�doi�10.1073�pnas.0403317101 Laakkonen et al.

Dow

nloa

ded

by g

uest

on

July

11,

202

0

DiscussionWe report here that LyP-1, a peptide that specifically binds totumor lymphatics and tumor cells, strongly accumulates in breast

cancer xenografts after an i.v. injection. The peptide and aclosely related variant of it preferentially localize in hypoxicareas within the tumors. We also show that systemically admin-istered LyP-1 causes tumor cell apoptosis, reduces the number oftumor lymphatics, and inhibits tumor growth in mice bearingbreast cancer xenografts. These results suggest that it may bepossible to develop LyP-1-based cancer therapies.

The LyP-1 peptide shows strong accumulation in the MDA-MB-435 tumors, including metastases from these tumors. Theefficacy and specificity of this peptide was sufficient to allow usto visualize orthotopic tumors in intact mice based on fluores-cence. Although fluorescence-based imaging of tumors formedby GFP-producing cells in intact animals is possible (19), achiev-ing it with an i.v. injected material may be unique. The remark-able tumor-homing efficiency of the LyP-1 peptide may bebecause of the propensity of this peptide to become internalizedby cells. Cells that bind the LyP-1 peptide transport it across thecell membrane, into the cytoplasm and the nucleus. In thisregard, LyP-1 is similar to the Tat peptide and other cell-penetrating peptides, which are also taken up by cells (24). Animportant difference is that our LyP-1 peptides are cell type-specific and deliver a payload to specific target cells: thelymphatic endothelial and tumor cells in tumors that display the‘‘receptor’’ for these peptides. The internalization is likely tocontribute to the effectiveness of these peptides in becomingconcentrated in the targeted tumors. If this efficacy can bereproduced in clinical settings, LyP-1-directed targeting of con-trast agents may become useful in tumor detection.

Fig. 4. LyP-1 peptide causes cell death in vitro and inhibits tumor growth in vivo. (A) LyP-1 (F, solid line) causes a dose-dependent release of lactatedehydrogenase from the cultured MDA-MB-435 cells, whereas a control peptide (CRVRTRSGC, E) has no effect. LyP-1 does not release lactate dehydrogenasefrom human C8161 melanoma cells (F, dotted line). (B) Mice bearing MDA-MB-435 tumors were injected twice a week with 60 �g of LyP-1 or its inactive variant(CGEKRTRGC), or with PBS. There were five mice�group; the treatment was started 4 weeks after the inoculation of the tumor cells (1–2 weeks after the tumorsbecame palpable) and lasted 4 weeks. One experiment of three is shown. LyP-1 reduced the mean tumor volume by an average of 50% (P � 0.05). (C and D) TUNELstaining (red) reveals clusters of apoptotic cells in the LyP-1-treated (C), but not control-treated (D), tumors. Blue, 4�,6-diamidino-2-phenylindole staining ofnuclei. Magnification, �200. The error bars in B show SEM.

Fig. 5. LyP-1 peptide reduces the number of tumor lymphatics. Tumorsections were stained with antibodies against CD-34 and podoplanin tovisualize and count tumor-associated blood vessels and lymphatics. LyP-1reduced the number of lymphatic vessels by an average of 85% (three exper-iments). Blood vessel density in the same tumors was affected less (averagereduction, 39%). The error bars show SD.

Laakkonen et al. PNAS � June 22, 2004 � vol. 101 � no. 25 � 9385

MED

ICA

LSC

IEN

CES

Dow

nloa

ded

by g

uest

on

July

11,

202

0

Our results show that treatment of tumor cells with the LyP-1peptide causes cell death. This effect is specific because cells thatdo not bind LyP-1 were not affected. The tumor cell apoptosiswe observed in vivo indicates that the LyP-1-binding cells die byapoptosis.

Whereas the mechanism whereby LyP-1 kills cells remains tobe elucidated, the proapoptotic effect seems to be directedagainst tumor cells that are under stress, as LyP-1 colocalizedwith a tissue hypoxia marker in vivo, and serum starvationenhanced LyP-1 binding and internalization by cultured tumorcells in vitro. It will be important to identify the molecule(receptor) to which LyP-1 binds at the cell surface (and that maymediate the proapoptotic effect of LyP-1). Our efforts to isolatea LyP-1 receptor by affinity chromatography and various cloningmethods have not yet been successful.

Treatment of tumor-bearing mice with the LyP-1 peptidesuppressed tumor growth. It also drastically reduced the expres-sion of lymphatic endothelial markers in the treated tumors. Thislatter result suggests that LyP-1 is also cytotoxic�proapoptoticfor lymphatic endothelial cells in tumors. As tumor lymphaticshave not been shown to be important for tumor growth (25), itis likely that the antitumor activity of LyP-1 is related to its effecton tumor cells rather than tumor lymphatics. However, given thedemonstrated role of tumor lymphatics in metastasis (15, 16, 22),destroying tumor lymphatics with LyP-1 may be particularly

effective in curtailing lymphatic spread of tumors. As lymphaticsappear to be the first target of LyP-1 in tumors (4), the antitumoreffect of LyP-1 may be particularly pronounced on tumor cellswithin and close to the lymphatics, which are likely to be the cellsmost probable to spread through the lymphatic system.

Hypoxia enhances metastasis (23, 26), and LyP-1 selectivelytargets tumor cells in the hypoxic areas of tumors. This may beanother pathway through which LyP-1 could suppress metasta-sis. MDA-MB-435 tumors are highly metastatic, and the VEGF-C-expressing cells are even more aggressive in that regard. In thisstudy, we evaluated the effects of LyP-1 on established primarytumors. As metastasis had already occurred at the time thetreatment began, we could not evaluate the effect of LyP-1 on themetastatic spread. Studies to determine the effects of LyP-1 onmetastasis are underway. Nonetheless, the data already at handdefine this peptide as a potentially unique tool for tumordiagnosis and treatment.

We thank Dr. Randall Johnson for reagents, Drs. Kari Alitalo and EvaEngvall for comments on the manuscript, and Roslind Varghese forediting. This work was supported by National Cancer Institute GrantCA82713, Department of Defense Grant DAMD 17-02-1-0315 (to E.R.),Cancer Center Support Grant CA30199, and National Cancer InstituteGrant CA099258-01 (to AntiCancer, Inc.). P.L. received support fromthe Academy of Finland and Biocentrum Helsinki. M.E.A. was sup-ported by Department of Defense Fellowship DAMD17-02-1-0308.

1. Ruoslahti, E. (2002) Nat. Rev. Cancer 2, 83–90.2. Hoffman, J. A., Giraudo, E., Singh, M., Zhang, L., Inoue, M., Porkka, K.,

Hanahan, D. & Ruoslahti, E. (2003) Cancer Cell 4, 383–391.3. Joyce, J. A., Laakkonen, P., Bernasconi, M., Bergers, G., Ruoslahti, E. &

Hanahan, D. (2003) Cancer Cell 4, 393–403.4. Laakkonen, P., Porkka, K., Hoffman, J. A. & Ruoslahti, E. (2002) Nat. Med.

8, 751–755.5. Breiteneder-Geleff, S., Soleiman, A., Kowalski, H., Horvat, R., Amann, G.,

Kriehuber, E., Diem, K., Weninger, W., Tschachler, E., Alitalo, K. & Kerjas-chki, D. (1999) Am. J. Pathol. 154, 385–394.

6. Banerji, S., Ni, J., Wang, S. X., Clasper, S., Su, J., Tammi, R., Jones, M. &Jackson, D. G. (1999) J. Cell Biol. 144, 789–801.

7. Wigle, J. T. & Oliver, G. (1999) Cell 98, 769–778.8. Achen, M. G., Jeltsch, M., Kukk, E., Makinen, T., Vitali, A., Wilks, A. F.,

Alitalo, K. & Stacker, S. A. (1998) Proc. Natl. Acad. Sci. USA 95, 548–553.9. Jeltsch, M., Kaipainen, A., Joukov, V., Meng, X., Lakso, M., Rauvala, H.,

Swartz, M., Fukumura, D., Jain, R. K. & Alitalo, K. (1997) Science 276,1423–1425.

10. Wigle, J. T., Harvey, N., Detmar, M., Lagutina, I., Grosveld, G., Gunn, M. D.,Jackson, D. G. & Oliver, G. (2002) EMBO J. 21, 1505–1513.

11. Jackson, D. G., Prevo, R., Clasper, S. & Banerji, S. (2001) Trends Immunol. 22,317–321.

12. Leu, A. J., Berk, D. A., Lymboussaki, A., Alitalo, K. & Jain, R. K. (2000) CancerRes. 60, 4324–4327.

13. Karpanen, T., Egeblad, M., Karkkainen, M. J., Kubo, H., Yla-Herttuala, S.,Jaattela, M. & Alitalo, K. (2001) Cancer Res. 61, 1786–1790.

14. Pepper, M. S. (2001) Clin. Cancer Res. 7, 462–468.

15. Stacker, S. A., Caesar, C., Baldwin, M. E., Thornton, G. E., Williams, R. A.,Prevo, R., Jackson, D. G., Nishikawa, S., Kubo, H. & Achen, M. G. (2001) Nat.Med. 7, 186–191.

16. Mandriota, S. J., Jussila, L., Jeltsch, M., Compagni, A., Baetens, D., Prevo, R.,Banerji, S., Huarte, J., Montesano, R., Jackson, D. G., et al. (2001) EMBO J.20, 672–682.

17. Valtola, R., Salven, P., Heikkila, P., Taipale, J., Joensuu, H., Rehn, M.,Pihlajaniemi, T., Weich, H., deWaal, R. & Alitalo, K. (1999) Am. J. Pathol. 154,1381–1390.

18. Saharinen, P., Tammela, T., Karkkainen, M. J. & Alitalo, K. (2004) TrendsImmunol., in press.

19. Yang, M., Baranov, E., Jiang, P., Sun, F. X., Li, X. M., Li, L., Hasegawa, S.,Bouvet, M., Al-Tuwaijri, M., Chishima, T., et al. (2000) Proc. Natl. Acad. Sci.USA 97, 1206–1211.

20. Lord, E. M., Harwell, L. & Koch, C. J. (1993) Cancer Res. 53, 5721–5726.21. Schueneman, A. J., Himmelfarb, E., Geng, L., Tan, J., Donnelly, E., Mendel,

D., McMahon, G. & Hallahan, D. E. (2003) Cancer Res. 63, 4009–4016.22. Skobe, M., Hawighorst, T., Jackson, D. G., Prevo, R., Janes, L., Velasco, P.,

Riccardi, L., Alitalo, K., Claffey, K. & Detmar, M. (2001) Nat. Med. 7, 192–198.23. Rofstad, E. K., Rasmussen, H., Galappathi, K., Mathiesen, B., Nilsen, K. &

Graff, B. A. (2002) Cancer Res. 62, 1847–1853.24. Lundberg, P. & Langel, U. (2003) J. Mol. Recognit. 16, 227–233.25. He, Y., Kozaki, K., Karpanen, T., Koshikawa, K., Yla-Herttuala, S., Takahashi,

T. & Alitalo, K. (2002) J. Natl. Cancer Inst. 94, 819–825.26. Zhong, H., De Marzo, A. M., Laughner, E., Lim, M., Hilton, D. A., Zagzag,

D., Buechler, P., Isaacs, W. B., Semenza, G. L. & Simons, J. W. (1999) CancerRes. 59, 5830–5835.

9386 � www.pnas.org�cgi�doi�10.1073�pnas.0403317101 Laakkonen et al.

Dow

nloa

ded

by g

uest

on

July

11,

202

0